For your application you need a circuit providing a much higher input impedance than this. This can only deliver a FET-input. 7...10MOhm is a much too low load impedance, when being connected to the probe. Especially, if you take the relevant temperature dependency of input impedance into consideration.



This is the circuit used by bitscope:

http://www.bitscope.com/design/ha...Bs11-5.PDF

This circuit is a simple Source-follower with a current source as source resistance. Source drain current is about 5mA, guaranteeing a rather low output impedance (in the 100Ohm range). So, a load impedance of 1MOhm at the FET-buffer output would only result in a gain error of less than 0.1%, which is totally negligible for a scope application. So, the AD844 is well suited to buffer the FET-buffer.



There are not many circuits being simpler that this Source-follower! Again, this circuit presents no gain but 1, so what would you expect to degrade? No, this circuit is highly stable and is widely used in all sectors of electronics. It benefits from the brilliant HF performance of FET.



Yes.



I designed the following:

5V (=50V total screen = 2Vpp at input of ADC)     - :100 x2 :5 x10
2V (=20V total screen = 2Vpp at input of ADC)     - :100 x2 :2 x10
1V (=10V total screen = 2Vpp at input of ADC)     - :100 x2 :1 x10
0.5V (=5V total screen = 2Vpp at input of ADC)    - :10  x2 :5 x10
0.2V (=2V total screen = 2Vpp at input of ADC)    - :10  x2 :2 x10
0.1V (=1V total screen = 2Vpp at input of ADC)    - :10  x2 :1 x10
0.05V (=0.5V total screen = 2Vpp at input of ADC) - :1   x2 :5 x10
0.02V (=0.2V total screen = 2Vpp at input of ADC) - :1   x2 :2 x10
0.01V (=0.1V total screen = 2Vpp at input of ADC) - :1   x2 :1 x10

In combination with a 1:1/10:1 probe you can handle an input voltage range of up to 500V.

More tomorrow.

Kai